Robust Deep Signed Graph Clustering via Weak Balance Theory

2025-02-08

AI Summary

This paper addresses two key challenges in signed graph clustering: poor noise robustness and cluster boundary shrinkage induced by strong balance theory (“the enemy of my enemy is my friend”). To this end, we propose DSGC, an end-to-end deep framework grounded in weak balance theory. DSGC is the first method to systematically integrate weak balance principles across the entire pipeline—preprocessing, augmentation, encoding, and optimization. Specifically, it introduces a Violation Sign-Refine module to denoise edges violating weak balance; employs a density-driven signed augmentation strategy to relax the restrictive strong balance assumption; designs a weak-balance-guided signed graph neural network; and incorporates a regularized clustering loss for joint optimization. Extensive experiments on synthetic and real-world datasets demonstrate that DSGC significantly improves clustering accuracy and noise robustness, consistently outperforming state-of-the-art methods and establishing a new benchmark.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Signed graph clustering is a critical technique for discovering community structures in graphs that exhibit both positive and negative relationships. We have identified two significant challenges in this domain: i) existing signed spectral methods are highly vulnerable to noise, which is prevalent in real-world scenarios; ii) the guiding principle ``an enemy of my enemy is my friend'', rooted in extit{Social Balance Theory}, often narrows or disrupts cluster boundaries in mainstream signed graph neural networks. Addressing these challenges, we propose the underline{D}eep underline{S}igned underline{G}raph underline{C}lustering framework (DSGC), which leverages extit{Weak Balance Theory} to enhance preprocessing and encoding for robust representation learning. First, DSGC introduces Violation Sign-Refine to denoise the signed network by correcting noisy edges with high-order neighbor information. Subsequently, Density-based Augmentation enhances semantic structures by adding positive edges within clusters and negative edges across clusters, following extit{Weak Balance} principles. The framework then utilizes extit{Weak Balance} principles to develop clustering-oriented signed neural networks to broaden cluster boundaries by emphasizing distinctions between negatively linked nodes. Finally, DSGC optimizes clustering assignments by minimizing a regularized clustering loss. Comprehensive experiments on synthetic and real-world datasets demonstrate DSGC consistently outperforms all baselines, establishing a new benchmark in signed graph clustering.

Problem

• Enhances noise resistance in signed graph clustering
• Broadens cluster boundaries using Weak Balance Theory
• Improves accuracy in community structure detection

Innovation

• Weak Balance Theory usage
• Violation Sign-Refine denoising
• Density-based Augmentation enhancement